The purpose of this proposal is to investigate the inter-relationships between bio-behavioral and environmental factors in adolescent tobacco use. A multilevel analysis of The National Longitudinal Study of Adolescent Health (Add Health) will assess multiple influences on adolescent smoking. The rationale underlying the multilevel approach is that no single factor from one level has been identified that completely accounts for adolescent tobacco use. Instead, a complex interaction between risk and protective factors from proximal and distal environments requires the adoption of a multi-faceted approach. The proposed study will examine the effect of state and school tobacco control policies on adolescent smoking behavior as moderated by parental smoking and parenting styles while controlling for individual demographic characteristics. The approach to achieve this goal is to specify and estimate a model combining bio-behavioral and environmental influences using a nationally-representative data set. Neither the bio-behavioral influences models nor the environmental influences models alone adequately explain why adolescents take up smoking and why they continue to smoke. The proposed study will use a novel approach in that it incorporates these two approaches in a multilevel analysis of smoking behavior in youth, which permits more precision in the estimates of the effects of the conjunction of various policies and factors by accurately accounting for nonindependence effects associated with each level of the model. Nationally representative data from Add Health, Waves I and III (a 5-year longitudinal survey) will be used to estimate a mixed-effects model with fixed effects and two-levels of random effects from state and school. The findings will inform tobacco control policy makers as to how to direct spending so as to maximize efficacy in terms of the prevention of smoking initiation in the short term and the promotion of cessation in general in the long term.